This week, I built and tested linear-model and neural-network ensembles to predict real-estate prices. Both had lower validation error than their members’ average, but neither beat its best individual model

Deep Learning
1